Classifications

• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04W—WIRELESS COMMUNICATION NETWORKS
  • H04W48/00—Access restriction; Network selection; Access point selection
  • H04W48/02—Access restriction performed under specific conditions
  • H04W48/06—Access restriction performed under specific conditions based on traffic conditions
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04W—WIRELESS COMMUNICATION NETWORKS
  • H04W28/00—Network traffic management; Network resource management
  • H04W28/16—Central resource management; Negotiation of resources or communication parameters, e.g. negotiating bandwidth or QoS [Quality of Service]
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04W—WIRELESS COMMUNICATION NETWORKS
  • H04W28/00—Network traffic management; Network resource management
  • H04W28/16—Central resource management; Negotiation of resources or communication parameters, e.g. negotiating bandwidth or QoS [Quality of Service]
  • H04W28/18—Negotiating wireless communication parameters
• • H—ELECTRICITY
  • H04—ELECTRIC COMMUNICATION TECHNIQUE
  • H04W—WIRELESS COMMUNICATION NETWORKS
  • H04W74/00—Wireless channel access

Definitions

• TITLE RELATION-BASED FUZZY- AND DISCRETE-LOGIC BASED MULTIDIMENSIONAL DECISION AND DYNAMIC TUNING CONTROL
• the present invention relates to policy-based decision- making in communications network elements .
• the invention has been developed primarily for use in Third Generation (3G) telecommunications networks, and more particularly for allowing admission control in an Internet Protocol (IP) Radio Access Network (RAN) element.
• IP Internet Protocol
• RAN Radio Access Network
• the invention has application in other policy-based decision-making elements that use rule-based logic for configuration and dynamic tuning, including (but not limited to) network elements, Policy Decision Points (PDPs) , and rule-based engines, whether acting alone or co-operating with other decision-making elements .
• GSR Guaranteed Bit-Rate
• QoS Quality of Service
• admission control denotes subscriber admission control in 3G .-networks implemented by standardized policy management functionality.
• the admission control has two purposes in a 3G network:
• MT Mobile Terminal
• UTRAN IMTS Terrestial Radion Access Network
• CN EDGE CN EDGE and Gateway
• admission control and QoS are based on different network load factor Key Performance Indicators (KPIs) , which can be measured within the network. These factors can include, for example, edge-to-edge delay, network load in the local network domain, available bandwidth in the domain, available radio channels and channel types, and Bit Error Rate (BER) . All of these values can be used to tune up the multidimensional admission control (MDAC) model.
• KPIs Key Performance Indicators
• MDAC multidimensional admission control
• the execution load in the admission control unit can be relatively high. It is an aim of embodiments of the present invention to provide an improved method and apparatus for implementing policy based decisions in a communications network. In a particularly preferred form, the aim of embodiments of the invention is to reduce the execution load associated with implementing admission control procedures in a 3G network.
• an admission control .” unit for users in a wireless communication system said unit being arranged to control the admission of calls arriving from the users depending on a parameter which is representative of the load in said system, wherein the load parameter is derived from a fuzzy logic composition of at least two indicators, each defining different performance characteristics of the load in said system.
• a method for controlling the admission of calls in a wireless communication network having a load comprising the step of: receiving at least two indicators each defining a different performance characteristic of the load in the network; combining said indicators using fuzzy logic to determine a parameter representative of the load of the network; and deciding based on said load parameter whether to admit calls arriving from the users .
• a wireless communication system having a load formed by calls transferred between users of the system, the system comprising means for controlling the admission of calls arriving from the users depending on a parameter which is representative of the system load, wherein the value of the load parameter is a fuzzy logic combination of at least two indicators each defining different performance characteristics of load in the system.
• Figure 1 shows admission control points in a 3G network, in which the invention is implemented
• Figure 2 shows an example of min-min and max-max interpretation of data, in accordance with the invention
• FIG. 3 shows a discrete Call Admission Control (CAC) embodiment of the invention
• Figure 4 is a visual representation of a network load decision made in accordance with an embodiment of the invention.
• FIG. 5 shows an implementation of an embodiment of the invention
• Figure 6 is a visual representation of a Multidimensional Admission Control (MDAC) structure, in accordance with an embodiment of the invention.
• MDAC Multidimensional Admission Control
• QoS control is based on correctly operating admission control (i.e. controlling user access) to offer adequate network resources for 3G QoS on IP and UMTS layers.
• admission control i.e. controlling user access
• the relationship between admission control and QoS is based on different network load factors (KPIs) which can be obtained from the network .
• Figure 6 depicts the basic idea of the multidimensionality of the QoS attributes (or KPIs) .
• All of the QoS attributes mentioned in the 3GPP [TS23.107] standard can be used as part of the dynamically tuned delay or load part or they can be split into separate dimensions in the admission control plane.
• TS23.107 is hereby incorporated by reference .
• Each of the KPIs are to some extent related to each other and they are not fully independent and orthogonal. Also, the correlation between KPIs varies in every environment, for example as a result of different configurations (or combination of devices), traffic mixes or network loads.
• fuzzy logic provides a more general definition than conventional Boolean logic.
• Specific systems have parameters that are defined as either being false “0” or true “1” which are - often referred to as “crisp” numbers, however by using a range of continuous values between "0” and “1” fuzzy logic is extended so as to incorporate the idea of partial truth.
• fuzzy subsets also known as “membership functions” which will be referred to hereinafter
• membership functions are another important characteristic of fuzzy logic, and allow values of a system to be better defined in terms of their partial truth.
• Membership functions are capable of taking many different forms and are often represented as triangles as shown in the waveforms of Figure 5 (which will be discussed later) , but can also be a simple straight line or more complex functions .
• membership functions and fuzzy logic can often be based on more than one characteristic (or subset) . So, in the example, it is possible to re-define a membership function to take into account both height and age so that a person can be judged on being "tall for their age” . This is often referred to as a fuzzy relation (or two-dimensional membership function) .
• a fuzzy system is defined by a collection of membership functions and rules (i.e. rule base) to reason about data.
• composition refers to the process when all of the fuzzy subsets assigned to a variable (or set) are combined together to form a single fuzzy subset.
• compositions are possible and embodiments of the present invention describe using max-* and max-min compositions.
• defuzzification is an optional process which can be used to convert from a fuzzy number to a crisp number .
• fuzzy logic is ideal for modelling complex real world systems and is therefore perfectly suited for admission control of telecommunication network having a plurality of potentially correlated KPIs. That is, fuzzy logic is able to take into account the correlation between each KPI or at least can be expressed as a fuzzy relation with a membership function that describes the extent that the KPIs in question are related to each other.
• the preferred embodiment is a method and corresponding apparatus for implementing a Multidimensional Admission Control (MDAC) in a 3G network.
• MDAC Multidimensional Admission Control
• ARP Allocation/Retention Priority
• the term multidimensionality refers to the incorporation of multiple measured network KPIs into one subscriber admission control load value, referred to herein as Mload KPI .
• the value of the Mload KPI can be any combination of available KPIs.
• multidimensional admission control is implemented in fuzzy logic using max-min or max-* composition.
• the multidimensionality is based on, for example, the following factors: edge-to-edge delay, BER, price factor and Mload KPI parameter.
• edge-to-edge delay BER
• price factor Mload KPI parameter.
• Each of these factors represents a measure of network quality at a particular time.
• the value of each factor represents a certain portion of resource load, e.g. network bandwidth, and overall delay budgeting and jitter.
• the MDAC of the preferred embodiment is based on classical fuzzy logic max-* or max-min fuzzy relation, as shown in equations 1 and 2: For max-*
• ⁇ 2 ( » max(min(/. ⁇
• x, y and z can represent any three dimension admission parameter combination.
• membership functions ⁇ R1 and ⁇ R2 are fuzzy values for the traffic load in the network domain.
• the Mload KPI parameter is used for a connection admission control decision.
• This relation deduction model can also be used in a non- fuzzy form. It can be done by removing membership functions ⁇ R1 and ⁇ R2 from formulas 1 and 2 and replacing them with discrete values obtained from the discrete functions F R ⁇ and F R2 in equations 3 and 4.
• Rx and __ 2 Relation tables constructed from functions F R1 and F R2 ; and x, y and z and are the parameters for functions F R1 and F R2 .
• the final value generated is a real measure (%) of the actual network load in the domain, which can be used in making an admission decision.
• MULTIDIMENSIONAL ADMISSION CONTROL DISCRETE LOGIC EXAMPLE 1 The next example depicts the functionality of the MDAC using two 2-D tables where the dimension y is the Mload KPI value and each of the other parameters can be any desired traffic property such as delay, Mean Opinion Score (MOS) , jitter etc. It will be appreciated that any suitable number of tables can be used.
• MOS Mean Opinion Score
• Parameters such as delay can also be fuzzy, as in the example.
• the goal is to determine what is the FinalLoad value for the call admission control (CAC) decision.
• CAC call admission control
• R i Selected(y) .60 .65 .70 .71
• the next step is to get the value of Mload KPI from the system and select the appropriate FinalLoad value from the max-* or max-min relation table. This means that if a measured Mload KPI value is 75%, the FinalLoad would be .97 in max-* case and .70 in max-min case.
• the fuzzy membership function value is a measure and value of association for the parameter to a measurement dimension. In this case the fuzzy function value has been directly translated into a network load value.
• Figure 2 depicts how the min and max methods should be interpreted.
• the decision system always chooses the path that produces the lowest Mload KPI value, whereas the max case is the opposite, in that the highest Mload KP ⁇ value is used. All of the parameter values can be fuzzy or discrete.
• FIG. 3 presents a simple practical example for determining a final load factor definition by discrete relation. Case parameters are: MOS 3.4 and Delay of 265 ms .
• predefined mapping tables Ri oad and R 2 oad are used. These tables may be achieved by RT-measurements (real time) from the network or by theoretical simulations. The tables can be short-term or long-term controlled as required and they act as the main tuning and configuration point to the CAC .
• the first phase is to measure or calculate domain MOS and Delay values for the first stage, R'lLoad and R'21-oad relation vector selections. All values can be fuzzy membership or discrete values.
• the load selection changes from the delay curve to the MOS curve as the load value changes from 70% to 80% whilst moving upwards along the curves.
• the situation will be the same with any other crossing property mapping case.
• the values in the tables are membership values requiring more computation effort to calculate than the pre-defined discrete values .
• the membership values can be calculated using a variety of so-called "T-norms", which are well known and are beyond the scope of the present invention. That is, in the fuzzy case the values also need to be defuzzified in order to get a final Mload KPI .
• a fuzzification process needs to be undergone as shown in Figure 5, in which the membership values of the KPIs, i.e. Delay and MOS, are fuzzified to form the fuzzified values c. ⁇ and ⁇ 2 respectively. Also, Mload KPI is represented as the ⁇ c (z) fuzzified value.
• the inferred consequence C can be calculated from equation 5.
• ⁇ c (z) ( ⁇ x ⁇ ⁇ c ⁇ (z)) v ( ⁇ 2 A ⁇ C2 (z)) ⁇
• Ruling can also be expressed for a common N rule case, equation 1 -.
• Wk is the value where the membership function Hk reaches its maximum (i.e. "1" if normalized) .
• the final admission control decision can then be made according to the following decision rules:
• All other subscriber classes will be handled in a similar way. Only the threshold values will be different, lower for lower class subscribers. There will also be a special class, which will bypass all the other classes. It uses an ARP value 0 and it is for government, official and emergency call priorities. In some cases the ARP value 0 can be used for both official and Gold subscriber classes .
• the rules used for the rule based admission control system can be in the following form.
• Mload KP! GoldDroppingTliresholdValue then Drop all GoldUserCallType and SilverUserCallType and BronzeUserCallType and EconomyUserCallType
• LPDP is the point where policy decisions are actually made whereas the PEP is the point where the policy descsions are actually enforced.
• policy based management is not mandatory with MDAC, it is envisaged in other embodiments that MDAC could supplement and cooperate nicely with policy based management standards.
• MDAC provides the FinalLoad KPI value to policies (rules) applied in PEP on the arriving calls from the subscribers to the network..-
• All QoS attribute parameter values mentioned in 3GPP [TS23.107] can be used as a part of the dynamically tuned delay or load part or they can be separate dimensions in the admission control plane.
• the present embodiments can be used for 3G network admission control development work, tuning of an existing
• the present invention in other embodiments can implement policy-based decision-making in any suitable element in a telecommunications network.

Landscapes

• Engineering & Computer Science (AREA)
• Computer Networks & Wireless Communication (AREA)
• Signal Processing (AREA)
• Computer Security & Cryptography (AREA)
• Quality & Reliability (AREA)
• Data Exchanges In Wide-Area Networks (AREA)
• Feedback Control In General (AREA)
• Logic Circuits (AREA)
• Mobile Radio Communication Systems (AREA)
• Executing Machine-Instructions (AREA)
• Error Detection And Correction (AREA)
• Hardware Redundancy (AREA)
• Computer And Data Communications (AREA)

Applications Claiming Priority (3)

Publications (2)

Family

ID=29712107

Family Applications (1)

Country Status (7)

Families Citing this family (4)

Family Cites Families (13)

• 2003
  • 2003-05-30 EP EP03732908A patent/EP1510098B1/de not_active Expired - Lifetime
  • 2003-05-30 CN CNB038125307A patent/CN100381010C/zh not_active Expired - Fee Related
  • 2003-05-30 AT AT03732908T patent/ATE445981T1/de not_active IP Right Cessation
  • 2003-05-30 US US10/448,471 patent/US20040042455A1/en not_active Abandoned
  • 2003-05-30 WO PCT/IB2003/002601 patent/WO2003103327A1/en not_active Ceased
  • 2003-05-30 AU AU2003239719A patent/AU2003239719A1/en not_active Abandoned
  • 2003-05-30 DE DE60329678T patent/DE60329678D1/de not_active Expired - Fee Related

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events